Three cheers to you for starting to pull together HTML enhancements!
On the subject of query forms, maybe I just don't understand where
Tim's old <QUERYFORM> ideas were heading, but we're working on an
approach that I like better. I hinted at it in a recent message
regarding variable substitution, here's more detail.
Define new tags for acquiring information from the user about an
object retrieval. A basic set of these tags is:
1.<ENTRYBLANK TYPE=type LENGTH=length DEFAULT=default VAR=lval>
Prompt </ENTRYBLANK>
2.<QUESTION TYPE=type DEFAULT=default VAR=lval> Prompt </QUESTION>
3.<CHOICE DEFAULT=default VAR=lval>
<ALTERNATIVE VAL=value1> Prompt1
...
<ALTERNATIVE VAL=valuen> Promptn
</CHOICE>
A GUI browser, for example, would turn the first kind of tags into
embedded entry widgets and do value checking based on the *type*
(e.g., string, bool, int, float). Tags of the second kind might
turn into checkboxes, where the type could be 2-state or 3-state.
In both cases the Prompt would be a label string. The GUI browser
could turn CHOICE constructs into radio buttons or listboxes,
depending on the number of ALTERNATIVES, the homgeneity of the
Prompt lengths, etc.
A tty/emacs browser could turn these tags into screen prompts,
y/n questions, menu lists, whatever.
Either way, the key is that these tags set variables, named by VAR
parameters, first to the default value and then to any input by the
user. These variables can then be substituted into URLs, thereby
dynamically creating URLs based on how the user fills in a form.
This scheme does require a URL variable substitution scheme like
I described in a recent message. We're working on an implementation
as mods to the HTML 2.0 library. Tim, would you be interested in
these mods?
Jay Weber
EIT